Interacting cytoplasmic loops of subunits a and c of Escherichia coli F₁F₀ ATP synthase gate H⁺ transport to the cytoplasm

Significance

H⁺ transport through the membrane-traversing F₁F₀ ATP synthase mechanically drives bond formation between ATP and P_i. For the enzyme in the inner membrane of Escherichia coli, we show here that the H⁺ transport pathway extends beyond the lipid bilayer into cytoplasmic loops connecting transmembrane helices in both subunits a and c of the transmembrane F₀ sector. On the basis of chemical cross-linking, we postulate that the multiple loop regions functioning in H⁺ transport pack into a single domain, which then interacts with the cytoplasmic surface of the H⁺ channel within F₀. The interaction of the loop regions with the H⁺ channel is proposed to specifically gate H⁺ release to the cytoplasmic side of the membrane during ATP synthesis.

Abstract

H⁺-transporting F₁F₀ ATP synthase catalyzes the synthesis of ATP via coupled rotary motors within F₀ and F₁. H⁺ transport at the subunit a–c interface in transmembranous F₀ drives rotation of a cylindrical c₁₀ oligomer within the membrane, which is coupled to rotation of subunit γ within the α₃β₃ sector of F₁ to mechanically drive ATP synthesis. F₁F₀ functions in a reversible manner, with ATP hydrolysis driving H⁺ transport. ATP-driven H⁺ transport in a select group of cysteine mutants in subunits a and c is inhibited after chelation of Ag⁺ and/or Cd⁺² with the substituted sulfhydryl groups. The H⁺ transport pathway mapped via these Ag⁺(Cd⁺²)-sensitive Cys extends from the transmembrane helices (TMHs) of subunits a and c into cytoplasmic loops connecting the TMHs, suggesting these loop regions could be involved in gating H⁺ release to the cytoplasm. Here, using select loop-region Cys from the single cytoplasmic loop of subunit c and multiple cytoplasmic loops of subunit a, we show that Cd⁺² directly inhibits passive H⁺ transport mediated by F₀ reconstituted in liposomes. Further, in extensions of previous studies, we show that the regions mediating passive H⁺ transport can be cross-linked to each other. We conclude that the loop-regions in subunits a and c that are implicated in H⁺ transport likely interact in a single structural domain, which then functions in gating H⁺ release to the cytoplasm.

The F₁F₀-ATP synthase of oxidative phosphorylation uses the energy of a transmembrane electrochemical gradient of H⁺ or Na⁺ to mechanically drive the synthesis of ATP via two coupled rotary motors in the F₁ and F₀ sectors of the enzyme (1). H⁺ transport through the transmembrane F₀ sector is coupled to ATP synthesis or hydrolysis in the F₁ sector at the surface of the membrane. Homologous ATP synthases are found in mitochondria, chloroplasts, and many bacteria. In Escherichia coli and other eubacteria, F₁ consists of five subunits in an α₃β₃γδε stoichiometry. F₀ is composed of three subunits in a likely ratio of a₁b₂c₁₀ in E. coli and Bacillus PS3 (2, 3) or a₁b₂c₁₁ in the Na⁺ translocating Ilyobacter tartaricus ATP synthase (1, 4) and may contain as many as 15 c subunits in other bacterial species (5). Subunit c spans the membrane as a hairpin of two α-helices, with the first transmembrane helix (TMH) on the inside and the second TMH on the outside of the c ring (1, 4). The binding of Na⁺ or H⁺ occurs at an essential, membrane-embedded Glu or Asp on cTMH2. High-resolution X-ray structures of both Na⁺- and H⁺-binding c-rings have revealed the details and variations in the cation binding sites (4–8). In the H⁺-translocating E. coli enzyme, Asp-61 at the center of cTMH2 is thought to undergo protonation and deprotonation, as each subunit of the c ring moves past the stationary subunit a. In the functioning enzyme, the rotation of the c ring is thought to be driven by H⁺ transport at the subunit a/c interface. Subunit γ physically binds to the cytoplasmic surface of the c-ring, which results in the coupling of c-ring rotation with rotation of subunit γ within the α₃β₃ hexamer of F₁ to mechanically drive ATP synthesis (1).

E. coli subunit a folds in the membrane with five TMHs and is thought to provide aqueous access channels to the H⁺-binding cAsp-61 residue (9, 10). Interaction of the conserved Arg-210 residue in aTMH4 with cTMH2 is thought to be critical during the deprotonation–protonation cycle of cAsp-61 (1, 11, 12). At this time, very limited biophysical or crystallographic information is available on the 3D arrangement of the TMHs in subunit a. TMHs 2–5 of subunit a pack in a four-helix bundle, which was initially defined by cross-linking (13), but now, such a bundle, packing at the periphery of the c-ring, has been viewed directly by high-resolution cryoelectron microscopy in the I. tartaricus enzyme (14). Previously published cross-linking experiments support the identification of aTMH4 and aTMH5 packing at the periphery of the c-ring and the identification of aTHM2 and aTMH3 as the other components of the four-helix bundle seen in these images (13, 15, 16). More recently, published cross-linking experiments identify the N-terminal α-helices of two b subunits, one of which packs at one surface of aTMH2 with close enough proximity to the c-ring to permit cross-linking (17). The other subunit b N-terminal helix packs on the opposite peripheral surface of aTMH2 in a position where it can also be cross-linked to aTMH3 (17). The last helix density shown in Hakulinen and colleagues (14) packs at the periphery of the c-ring next to aTMH5 and is very likely to be aTMH1.

The aqueous accessibility of Cys residues introduced into the five TMHs of subunit a has been probed on the basis of their reactivity with and inhibitory effects of Ag⁺ and other thiolate-reactive agents (18–20). Two regions of aqueous access were found with distinctly different properties. One region in TMH4, extending from Asn-214 and Arg-210 at the center of the membrane to the cytoplasmic surface, contains Cys substitutions that are sensitive to inhibition by both N-ethylmaleimide (NEM) and Ag⁺ (18–20; Fig. 1). These NEM- and Ag⁺-sensitive residues in TMH4 pack at or near the peripheral face and cytoplasmic side of the modeled four-helix bundle (11, 13). A second set of Ag⁺-sensitive substitutions in subunit a mapped to the opposite face and periplasmic side of aTMH4 (18, 19), and Ag⁺-sensitive substitutions were also found in TMHs 2, 3, and 5, where they extend from the center of the membrane to the periplasmic surface (19, 20). The Ag⁺-sensitive substitutions on the periplasmic side of TMHs 2–5 cluster at the interior of the four-helix bundle predicted by cross-linking and could interact to form a continuous aqueous pathway extending from the periplasmic surface to the central region of the lipid bilayer (11, 13, 19, 20). We have proposed that the movement of H⁺ from the periplasmic half-channel and binding to the single ionized Asp61 in the c-ring is mediated by a swiveling of TMHs at the a–c subunit interface (16, 21–24). This gating is thought to be coupled with ionization of a protonated cAsp61 in the adjacent subunit of the c-ring and with release of the H⁺ into the cytoplasmic half-channel at the subunit a–c interface. The route of aqueous access to the cytoplasmic side of the c subunit packing at the a–c interface has also been mapped by the chemical probing of Cys substitutions and, more recently, by molecular dynamics simulations (22, 25, 26).

Fig. 1.

The predicted topology of subunit a in the E. coli inner membrane. The location of the most Ag⁺-sensitive Cys substitutions are highlighted in red (>85% inhibition) or orange (66–85% inhibition). The five proposed TMHs are shown in boxes, each with a span of 21 amino acids, which is the minimum length required to span the hydrophobic core of a lipid bilayer. The α-helical segments shown in loops 1–2 and 3–4 are consistent with the predictions of TALOS, based on backbone chemical shifts seen by NMR (29). Others have also predicted extensive α-helical regions in these loops (12, 30), but the possible positions remain largely speculative. aArg210 is highlighted in green. Figure is modified from those shown previously (21, 23, 24, 27).

We have also reported Ag⁺-sensitive Cys substitutions in two cytoplasmic loops of subunit a (27) and, more recently, in the cytoplasmic loop of subunit c (28). The mechanism by which Ag⁺ inhibited F₁F₀-mediated H⁺ transport was uncertain. Several of these substitutions were also sensitive to inhibition by Cd⁺², and these substitutions provided a means of testing whether Cd⁺² directly inhibits passive H⁺ transport through F₀ (28). In the case of two subunit c loop substitutions, Cd⁺² was shown to directly inhibit passive F₀-mediated transport activity. In this study, we have extended the survey to Cd⁺²-sensitive Cys substitutions in cytoplasmic loops of subunit a. We report four loop substitutions in which Cd⁺² inhibits passive F₀-mediated H⁺ transport. Further, in two cases, we show cross-linking between pairs of Cys substitutions, which lie in subunits a and c, respectively, and which individually mediate passive H⁺ transport activity. These results suggest that the a and c loops, which gate H⁺ release to the cytoplasm, fold into a single domain at the surface of F₀.

Results

Cytoplasmic Loop Cys in Subunit a Showing Reversible Ag⁺ or Cd⁺² Inhibition.

In Steed and Fillingame (28), Ag⁺-sensitive Cys substitutions in the cytoplasmic loop of subunit c were initially distinguished from each other according to the reversibility of Ag⁺ inhibition of F₁F₀-driven H⁺ transport by subsequent treatment with DTT. In one group of Ag⁺-sensitive mutants, Ag⁺ inhibition was caused by uncoupling F₁ from F₀, with frequent disruption of F₁ binding to F₀ and loss of F₁ from the membrane. In these cases, the Ag⁺ inhibition was not significantly reversed by DTT treatment. In a second group, Ag⁺ inhibition was completely reversed by DTT treatment. This reversal was postulated to be a result of the removal of blockage of H⁺ transport. For a subset of this second group of Cys substitutions, Cd⁺² was shown to directly block passive H⁺ transport mediated by purified F₀ reconstituted into liposomes. The same approach was used in this study to screen Ag⁺-sensitive Cys substitutions initially, and then Cd⁺²-sensitive Cys substitutions in the loops of subunit a. The smaller group of Cd⁺²-sensitive substitutions showing DTT-reversibility was then directly tested in an assay for passive H⁺ transport mediated by purified F₀ reconstituted into liposomes.

DTT reversibility tests for Ag⁺ inhibition for Cys substitutions that extend into the cytoplasmic 1–2 loop (aM93C), the 3–4 loop (aL195C, aV198C, and aS202C), or the C-terminal cytoplasmic tail (aY263C) of subunit a are shown in Fig. 2. Reversibility was expected if inhibition was caused by direct blockage of the proton channel through F₀, rather than disruption of the F₁ interaction with F₀. In all five cases, Ag⁺ inhibition of ATP-driven H⁺ pumping was reversed rapidly by DTT. The aM93C, aL195C, and aV198C substitutions are sensitive to inhibition by Ag⁺, but not Cd⁺². The aS202C and aY263C substitutions are also sensitive to Cd⁺² inhibition, and the inhibition of ATP-driven H⁺ transport by Cd⁺² proved to be rapidly reversed by DTT (Fig. 3). In addition, Fig. 3 shows that inhibition of ATP-driven H⁺ pumping by the Cd⁺²-sensitive aS199C and aK203C substitutions was rapidly reversed by DTT treatment. In sum, seven of the loop substitutions extending into the cytoplasmic region beyond the hydrophobic phase of the lipid bilayer show metal inhibition that is reversed by DTT. In experiments that follow, we show that the metal inhibition in at least three of these cytoplasmic loop substitutions can be directly attributed to blockage of H⁺ transport through the proton channel of F₀.

Fig. 2.

Reversal of Ag⁺-inhibition of ATP-driven H⁺ transport by DTT in wild type (A) and Cys-substituted subunit a mutants (specified in B–F). Inner membrane vesicles were incubated with 0.3 µg/mL ACMA and treated or not treated with 40 µM AgNO₃ before the addition of 1 mM ATP (final concentration) to initiate the fluorescence quenching reaction. Subsequently, DTT (2 mM final concentration) was added to the Ag⁺-treated membranes to reverse the inhibition. Quenching was terminated by addition of 0.5 µg/mL nigericin at the arrow marked N.

Fig. 3.

Reversal of Cd⁺²-inhibition of ATP-driven H⁺ transport by DTT in wild type (A) and Cys-substituted subunit a mutants (specified in B–E). Inner membrane vesicles were incubated with 0.3 µg/mL ACMA and treated or not treated with 300 µM CdCl₂ before the addition of 1 mM ATP (final concentration) to initiate the fluorescence quenching reaction. Subsequently, DTT (2 mM final concentration) was added to the Cd⁺²-treated membranes to reverse the inhibition. The quenching response was terminated by the addition of nigericin to 0.5 µg/mL at the arrow marked N, as illustrated with the samples not treated with Cd⁺².

Cd⁺² Inhibition of Passive H⁺ Translocation Through the Proton Channel of F₀.

We have developed an assay for measuring passive H⁺ translocation through F₀ reconstituted into K⁺-loaded liposomes in response to a valinomycin-induced K⁺-diffusion potential (28). Ag⁺ cannot be tested for inhibition of passive H⁺ transport in this assay because it complexes with valinomycin and abolishes the signal generated on valinomycin addition. In contrast, inhibition by Cd⁺² can be tested because of the lack of Cd⁺² interaction with valinomycin. We chose to compare the passive H⁺ transport activity of three Cys substitutions in the cytoplasmic 3–4 loop, and the aY263C substitution in the cytoplasmic C-terminal tail in passive H⁺ transport activity with the aS206C substitution, which is positioned toward the cytoplasmic end of aTMH4. ATP-driven H⁺ transport by aS206C membrane vesicles is inhibited by Ag⁺, Cd⁺², and NEM in a DTT-reversible manner (23). We therefore expected that the aS206C substitution would be active in passive H⁺ transport through F₀ and that Cd⁺² would inhibit the passive H⁺ transport. K⁺-diffusion potential-driven H⁺ transport into liposomes reconstituted with aS206C F₀ proved to be somewhat slower than that seen with liposomes containing wild-type F₀ (Fig. 4). Importantly, the activity observed with S206C liposomes was progressively inhibited by Cd⁺² additions over the range of 10–300 µM, whereas the same concentrations of Cd⁺² had much smaller effects on the H⁺ transport activity of liposomes reconstituted with wild-type F₀. The Cd⁺² sensitivity is consistent with a model in which the aS206C cysteine side-chain lies within the H⁺ transport pathway, such that Cd⁺²-Cys chelation can block proton movement through F₀.

Fig. 4.

Membrane potential-driven, passive H⁺ transport using liposomes reconstituted with wild type (A) and aS206C (B). Liposomes reconstituted with F₀ were loaded with K₂SO₄ and 5 µL of liposomes diluted into 3.2 mL HMN-SO₄ buffer lacking K⁺. ACMA was added to 0.3 µg/mL, and the fluorescence quenching response was initiated by the addition of valinomycin to 14 ng/mL at the arrow marked V to generate the K⁺ diffusion potential. CdSO₄ was added at the concentrations indicated before addition of valinomycin. The response was terminated by addition of nigericin to 0.5 µg/mL at the arrow marked N.

After reconstitution into liposomes, the F₀ from the aS199C, aS202C, and aK203C substitutions in the 3–4 cytoplasmic loop all showed H⁺ transport activity similar to that seen with the aS206C substitution, albeit somewhat slower (Fig. 5). Cd⁺² progressively inhibited the H⁺ transport activity of aS199C and aS202C F₀ over the range of 10–300 µM in a manner similar to that for aS206C F₀. In contrast, the aK203C substitution proved to be markedly less sensitive to Cd⁺², with significant inhibition only being observed at 300 µM. The H⁺ transport activity of liposomes reconstituted with aY263C F₀ was less than that of the other substitutions shown in Fig. 5, but the activity was inhibited by Cd⁺² additions over the range of 10–300 µM in a manner similar to the aS206C, aS199C, and aS202C substitutions. We think that the relatively low activity of aY263C F₀ in reconstituted liposomes, which was observed in two separate trials, may be a reflection of its instability during detergent solubilization and reconstitution. Note that the ATP-driven H⁺ transport activity of aY263C inner membrane vesicles was equivalent to that of the other substitutions shown in Fig. 3, which were then tested in the passive H⁺ transport experiments shown in Fig. 5. In this regard, for the aM215C, aI249C, and aQ252C transmembrane substitutions, our reconstitution attempts resulted in F₀ liposomes showing small quenching responses (in the range of 11–25%), despite the robust activity seen in ATP-driven H⁺ transport activity, using inner membrane vesicles from these substitutions. We again suspect this may be a reflection of the lack of stability of these F₀ substitutions during reconstitution.

Fig. 5.

Membrane potential-driven, passive H⁺ transport using liposomes reconstituted with F₀ from wild type (A) and Cys-substituted subunit a mutants (specified in B–E). Liposomes were reconstituted and the passive H⁺ transport assay carried out as described in Fig. 4.

Cross-Linking of H⁺ Transporting Regions in Loops of Subunit a and Subunit c.

Moore and colleagues (27) were the first to identify subunit a Cys substitutions in both the 1–2 loop (residues 86 and 93) and the 3–4 loop (residues 177, 195, 198, 201, 202, and 203), in which ATP-driven H⁺ transport was blocked by Ag⁺ treatment. Further, after introduction of pairs of Cys into the two loop regions, cross-linking between the Ag⁺-sensitive loop regions was demonstrated (27). The latter result suggested these regions of cytoplasmic loops might pack into a single domain that would gate H⁺ translocation from the middle of the membrane to the cytoplasm. Subsequently, Steed and Fillingame (28) demonstrated that the Cd⁺² sensitivity of the cI46C and cL48C substitutions in the cytoplasmic loop of subunit c could be attributed to passive F_o-mediated H⁺ transport and suggested that these and neighboring residues likely participated in gating H⁺ transport to the cytoplasm. The experiments presented here suggests that loop Cys substitutions in subunit a may also directly participate in passive H⁺ transport, according to their DTT reversibility and direct H⁺ transport assays. These findings led us to test whether the cytoplasmic loop regions of subunits a and c, which were implicated in H⁺ translocation function, could be cross-linked to each other. We attempted to cross-link cI46C and cL48C in subunit c to aM93C in the 1–2 loop or to aL195C, aS199C, and aK203C in the 3–4 loop of subunit a, using two bis methanethiosulfonate (MTS) reagents: 1,2-ethanediyl bis-MTS (M2M) and 1,4-butanediyl bis-MTS (M4M). Mutant subunit a carrying the aM93C substitution was cross-linked to subunit c encoding either the cI46C or cL48C substitutions, using either M2M or M4M (Fig. 6). In contrast, we were unable to detect cross-linking between the cI46C and cL48C region in subunit c and Cys at positions 195, 199, or 203 in the 3–4 loop of subunit a. We conclude that the putative proton gating regions encoding cI46C and cL48C and aM93C do pack closely to each other in the F₀ complex and probably interact in a single domain involved in gating H⁺ exit from the membrane interior to the cytoplasm. Further, on the basis of the previous cross-linking experiments described by Moore and colleagues (27), we suggest that the hypothesized gating region of the 3–4 loop, centered in the region of the aS199C and aS202C substitutions, also packs within a single domain gating H⁺ release to the cytoplasm (Discussion).

Fig. 6.

Cross-linking between aM93C in subunit a and cI46C or cL48C in subunit c. Inverted membrane vesicles containing the aM93C/cI46C or aM93C/cL48C double mutations were treated with 1.8 mM M2M or M4M, as described in Methods. After M2M or M4M treatment, DTT was added to some samples to test the reversibility of the cross-link formation. Cross-linked products were detected by immunoblotting with an antibody against subunit a.

Discussion

We previously identified multiple Cys substitution sites in cytoplasmic loops of subunit a, where ATP-driven H⁺ transport was inhibited by treatment with Ag⁺. Those findings raised the question of whether the loop regions identified were actual parts of the transport pathway gating H⁺ release to the cytoplasm. In similar Ag⁺-inhibition studies with Cys substitutions in the cytoplasmic loop of subunit c, we were able to distinguish two modes of F₁F₀ inactivation by Ag⁺, according to reversibility of inhibition by DTT (28). Inhibition was not reversible in mutants, where Ag⁺ disrupted the interaction of F₁ with F₀, often with a resultant disassociation of F₁ from the membrane. In a second class of substitutions, including Cys at positions 44, 45, 46, and 48, inhibition was fully reversed by DTT treatment. Two of these substitutions were also inhibited by Cd⁺², and in these cases (cI46C and cL48C), we were able to show that Cd⁺² directly inhibited passive H⁺ transport through the F₀ moiety of the ATP synthase complex. Using these criteria for distinguishing the mechanism of inhibition, we have further characterized seven Cys substitutions in cytoplasmic loop domains of subunit a. DTT was shown to reverse Ag⁺ and/or Cd⁺² inhibition for M93C in the 1–2 loop, several substitutions in the 3–4 loop (aL195C, aV198C, aS199C, aS202C, and aK203C), and the aY263C substitution in the C-terminal tail. Further, for three of the four Cd⁺² sensitive substitutions (aS199C, aS202C, and aY263C), passive F₀ mediated H⁺ transport into liposomes was strongly inhibited by Cd⁺². Passive H⁺ transport by aK203C liposomes was not as sensitive to Cd⁺², and Cd⁺² may inhibit ATP-driven H⁺ transport in this mutant by additional mechanisms.

Ag⁺-sensitive Cys in the 1–2 loop had previously been shown to cross-link with Ag⁺-sensitive Cys in the 3–4 loop and suggested that the possible H⁺-transporting regions might interact structurally in a single gating domain (27) (Fig. 7). Here we report experiments in which we attempted to cross-link the cytoplasmic loop cI46C and cL48C substitutions, which appear to lie within a H⁺ gating pathway to the cytoplasm, with cytoplasmic loop Cys in subunit a that are also implicated in H⁺ transport and release to the cytoplasm. We were able to cross-link both cI46C and cL48C to aM93C in the 1–2 loop of subunit a and suggest all three residues likely fold within a common domain gating H⁺ release to the cytoplasm. We failed to find cross-links between cI46C or cL48C and Cys substitutions at positions 195, 199, or 203 in the 3–4 loop of subunit a. However, according to a previous study showing cross-linking between both aV86C and aM93C in the 1–2 loop and aL195C in the 3–4 loop (27), we think it is likely that these two loop domains interact during H⁺ gating to the cytoplasm and that the transport pathway extends to the edge of the lipid bilayer, as shown in Fig. 7, and includes residues aS199C and aS202C; that is, substitutions that show Cd⁺²-sensitive, F₀-mediated, passive H⁺ transport (Fig. 5). Ishmukhametov and colleagues (30) have previously suggested a possible interaction between the cAsp44–cArg50 region in the subunit c-loop and the aGlu196 region of subunit a, which supports the interacting domain proposal suggested here. The previously observed cross-link between aV86C in loop 1–2 and the Ag⁺-sensitive aT179C at the center of loop 3–4 suggests the region around residue 179 may also fold into this domain.

Fig. 7.

Model of interacting regions between Ag⁺-sensitive Cys substitutions in the cytoplasmic loops of subunits a and c. The cross-links between Cys substitutions in loop regions of subunit a were described in Moore and colleagues (27). The cross-links between cI46C and cL48C and aM93C are described in Fig. 6. Cd⁺²-sensitive passive H⁺ transport was directly demonstrated for Cys substitutions at residues 46 and 48 of subunit c (28) and positions 199 and 202 of the subunit a 3–4 loop in experiments described in Fig. 5.

We previously reported an aqueous access pathway extending from the cytoplasm to the G58C in the center of cTMH2 that permitted access and modification by both small cations, such as Ag⁺ and Cd⁺², and also larger reagents, such as NEM and MTS-ethyl-trimethylammonium (22, 25). If such an aqueous pathway is to specifically conduct protons to the region of cGly58 and cAsp61, additional residues that specifically gate H⁺ access or egress would be required at the cytoplasm surface. We suggest that these H⁺-gating residues may reside in the cytoplasmic loop domains of subunits a and c, which are implicated by the experiments presented here in passive H⁺ transport to and from the cytoplasm. H⁺ gating by cytoplasmic loops would also provide an explanation for the electron cryomicroscopy images of TMHs in the related Thermus thermophilus ATP synthase. TMHs that are presumed to be equivalent to those at the subunit a–c interface in E. coli were suggested to splay apart as they approached the cytoplasmic surface (31). The presumed aqueous cavity extending into the lipid bilayer would obviously not specifically conduct protons, but if H⁺-gating residues were present in the cytoplasmic loops, which would not be detected in the electron cryomicroscopy images, the cavity could be made proton-specific, as we suggest here for the E. coli enzyme.

Methods

E. coli Strains.

All of the single Cys substitutions in subunits a and c were expressed in plasmid pCMA113, which encodes a Cys-less F₁F₀ complex (18). The mutant plasmids described previously were transferred into the chromosomal atp operon deletion strain JWP292 for biochemical characterization (27, 28). To purify mutant F₁F₀ complexes, the Cys substitutions were excised from the pCMA113 derivative plasmid and transferred into plasmid pFV2 (30). Similar to pCMA113, pFV2 codes a Cys-less F₁F₀ complex and a His tag is attached at the N terminus of subunit β to facilitate purification (32). The pFV2 derivatives were transformed into the atp deletion strain DK8 (33), and transformant cells were grown to provide membranes from which F₁F₀ was purified.

ATP-Driven Fluorescence Quenching of 9-Amino-6-Chloro-2-Methoxyacridine.

Inside-out membrane vesicles were prepared from JWP292 transformant strains and ATP-driven quenching of 9-amino-6-chloro-2-methoxyacridine (ACMA) fluorescence, determined as described previously (27, 28). Treatments with Ag⁺ or Cd²⁺ were carried out in HMK-NO₃ buffer [10 mM Hepes-KOH, 1 mM Mg(NO₃)₂, 10 mM KNO₃ at pH 7.5] or HMK-Cl (10 mM Hepes-KOH, 5 mM MgCl₂, 300 mM KCl at pH 7.5 buffer), respectively. To determine whether any of the observed metal-related inhibition of quenching was reversible, 1/100 volume of DTT was added to a final concentration of 2 mM after the initiation of quenching with ATP.

Purification of F₁F₀ and Reconstitution of F₀ in Liposomes.

The His-tagged F₁F₀ complex was purified as previously described (28), using the procedure originally described by Ishmukhametov and colleagues (32). Liposomes were prepared by the method of Wiedenmann and colleagues (34), with the modifications described previously (28). F₁F₀ liposomes were either used immediately for the preparation of K⁺-loaded F₀ liposomes, as described (28), or were frozen in liquid nitrogen for later determination of ATP-driven H⁺-pumping activity (28).

ΔΨ-Driven ACMA Fluorescence Quenching.

A 5-mL aliquot of K⁺-loaded F₀ liposomes (0.6 mg lipid) was added to 3.2 mL of stirring HMN-SO₄ buffer (10 mM Hepes-NaOH, 2.5 mM MgSO₄, 25 mM Na₂SO₄ at pH 7.5) supplemented with 0.3 µg/mL ACMA in a fluorimeter cuvette. Quenching of fluorescence was initiated by the addition of valinomycin to 14 ng/mL and terminated by the addition of nigericin to 0.5 μg/mL For Cd²⁺ treatment, CdSO₄ was added at the indicated concentration after the addition of liposomes and ∼20 s before the initiation of quenching with valinomycin.

Cys-Cys Cross-Linking.

Membranes from Cys-substituted mutants were washed with TMG-acetate buffer (50 mM Tris⋅acetate, 5 mM MgCl₂, 10% (wt/vol) glycerol at pH 7.5) to remove residual DTT from the membrane preparation buffer. Membranes were suspended in TMG-acetate buffer and treated with bis-MTS reagents to test for cross-linking. M2M and M4M were dissolved in dimethyl sulfoxide immediately before use and added to a final concentration of 1.8 mM. After 10 min at room temperature, 5 μL 0.5 M ethylenediamine tetraacetic acid (EDTA) was added and the membrane suspensions were incubated an additional 15 min at room temperature. For some samples, DTT was added to 20 mM to reduce the cross-link. All samples were then diluted with one volume 2× SDS sample buffer [0.125 M Tris⋅Cl, 20% (wt/vol) glycerol, 4% (wt/vol) SDS at pH 6.8] and incubated for 1 h at room temperature. SDS polyacrylamide gel electrophoresis and immunoblotting were carried out as previously described (16).